Re: Chunk reported full, while it shouldn't ...

"malcolm weallans" <malcolm.iiug@btopenworld.com> wrote in message news:c4f3e8$v2q$1@terabinaries.xmission.com... > > It's funny but a few weeks ago I saw on this newsgroup a comment to the > effect that training courses tell you about setting extent sizes > correctly, etc but that it isn't really necessary. And today I have > been trying to cope with a database that ran out of space (one table has > 194 extents!) and now this error. I think we should all pat lots of > attention to the need for space planning on IDS systems. I wrote the > original user instructions on this back in 1988 (Turbo in those days) > and many of them still hold. But the important thing is that all DBAs > should plan their disk space, set up extent sizes, and montior growth. > Hopefully if they follow that model they won't encounter problems like > this. You know that idiot who wrote that you don't need to bother with table planning, and your time is far better spent on analysing key processes? That irritating man who sniffs at all the great theory that's packed into the manual and the course notes and that helps you to get the sizing right for optimal performance. You know the guy, the one who arrogantly says that, beyond a cursory look at it all and subsequent second extent adjustment where necessary, it's all irrelevtant? Do you know him? He's me. I posted that post. Although I've attended couses (on Adabas/Natural) about planning out table and extent sizes, I've never, ever bothered in real life. And I stand by this. You just wing it on on the basis that there are three table sizes: small, medium and large. Anyone who sits around spending hours and hours on detailed table size planning simply doesn't have enough football to watch. Disk space is cheap now, Malc!